JPH067043B2 - Position detection method - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a position detection method, and more particularly to a vibration type position detection method having an automatic gain adjustment function.

[Technical background of the invention and its problems]

The vibration-type position detection method is used in photoelectric microscopes developed for reading scale lines such as reference scales built into precision measuring materials.Since it has high detection sensitivity, it is used today for various automatic positioning of semiconductor manufacturing equipment, etc. It has been used as a detection method for devices.

The principle of the vibration-type position detection method will be briefly described with an example when it is used in a photoelectric microscope. The basic configuration is as shown in FIG. 1, in which 1 is a light source, 2 is a focusing lens, 3 is a half mirror, 4 is an objective lens, 5 is an object to be inspected on which a mark 6 is formed, and 7 is a slit plate. , 8 is a photoelectric converter, 9 is a preamplifier, 10 is a synchronous detection circuit, 11 is an oscillator, 12
Is a vibrator, and 13 is an indicating meter. The scanning mechanism uses a slit or a vibrator 12 that vibrates an image, and the output signal of the photoelectric converter 8 has a different signal waveform depending on the mark position because the slit vibrates. When the displacement of the signal waveform is shown with respect to the position x of the mark, FIG.
It becomes like (i). As can be seen from FIG. 2, the signal waveform is a synchronizing signal with a frequency equal to the vibration frequency, so when the mark position coincides with the vibration center position of the slit, that is, at point (e), the signal has a double frequency. , The fundamental frequency component becomes zero. The vibration type photoelectric microscope is for electrically detecting this position. Further, the right diagram shows how the fundamental frequency component a _w changes with respect to the change of the mark position x. The method of changing the fundamental frequency component (output characteristic curve) changes depending on the mark width, slit width, vibration amplitude, etc., but it is zero at point (e) where the mark position coincides with the slit vibration center, and the sign before and after that point. Since it rotates in the reverse direction, it is suitable for zero meter detection and servo system drive. Note that extracting only the fundamental frequency component from the photoelectric signal and erasing the double frequency component is
It is easily realized by the synchronous detection circuit 10.

By the way, the shape of the output characteristic curve is the mark width, contrast,
Since it varies depending on the scanning (vibration) amplitude, the brightness of illumination, and the like, the relationship between the output and the mark position is not constant, and therefore it is necessary to perform calibration each time to obtain correct position information. If this is not done, the position setting will be inaccurate and it will take a lot of time. For example, when this position detection method is used in a semiconductor manufacturing apparatus typified by an optical exposure apparatus, since exposure is repeated through a large number of various processing steps, the mark contrast is a step in position detection before exposure. And the peak values P ₁ , P ₁ ′, P ₁ ″ of the curves of the outputs a _f , a _f ′, a _f ″ when the processes are different as shown in FIG. The inclination is different. Therefore, a method is considered in which information on the contrast of the mark due to the difference in the process is entered in a computer in advance to correct the output.
However, in this method, since the information on the contrast of the mark is based on prediction and cannot be completely grasped, accurate correction cannot be performed. Furthermore, since this output becomes unstable due to fluctuations in illuminance, the peak value and slope of the output change over time if sufficient stability of illuminance cannot be obtained. Recently, an electron beam transfer apparatus, which is considered to be promising for transferring submicron-order fine patterns, has been subjected to alignment applying this position detection principle, but in this case, a similar problem occurs.
Moreover, in this case, the electron beam is irradiated instead of irradiating the mark with light, but there is a phenomenon that the photocathode, which is the generation source, deteriorates with time, and the irradiation amount of the electron beam changes with time. Therefore, the output signal also changes with time. Since such changes over time are difficult to predict, it is difficult to accurately correct the output by prediction.

For this reason, it has become necessary to always keep the slope of the output characteristic curve constant by AGC (auto gain control) of the signal. As a method of AGC, it is easy to always bring the mark closer from one direction and control the gain so that the peak voltage of the output characteristic curve becomes a constant value. However, with this method, it is necessary to perform detection while scanning a certain time mark, which leads to a delay in the position detection time, leading to a reduction in the throughput of the device. Moreover, it is not possible to deal with the change of the signal over time while detecting the peak value of the output curve, and it becomes difficult to perform precise positioning in a short time.

On the other hand, as a method of AGC performed in real time, a method of detecting an effective value or a peak value of a signal and adjusting a gain so that the value is always constant is considered. However, the detection signal in the vibration-type position detection contains not only the fundamental wave component but also higher-order frequency components, and their output components are greatly different depending on the mark position, resulting in a complicated signal waveform. It was difficult to perform AGC with high accuracy without losing the position information.

[Object of the Invention]

An object of the present invention is to provide a vibration type that has the effect of automatic gain adjustment such that the position detection output characteristic becomes constant even if the input signal level changes, and that can perform position detection without impairing detection time or detection accuracy. It is to provide a position detecting method.

[Outline of Invention]

The essence of the present invention is that the fundamental wave component a obtained by synchronous detection is
_A specific function conversion for a _f and a _2f without directly using _f and the nth harmonic component (n is an integer of 2 or more), for example, the second harmonic component a _2f as a position detection signal. And signals a _f ^* , a _2f that are not affected by fluctuations in input level
^* _Is obtained, and these signals a _f ^* and a _2f ^* are used as position detection signals.

The output characteristics of the signals a _f and a _2f used as the position detection signals have the relationship shown in FIG. That, a _f zero mark position in terms of zero (coincident with the oscillation center),
a _2f indicates the peak value, and a _f indicates the peak value.
_2f indicates zero. And, it is a monotonous curve between the two peak values of a _f . These relationships are specific in the position detection of this type, a _f, a is obtained by some transformation of a _{2f f} ^*, which combination of a _2f ^* output characteristics satisfy the same relationship If,
a _f ^*, it can be used as a _2f ^* also still position detection signal. The present inventors have result of extensive research by focusing on this point, a _f ^* that satisfies the relationship described above, the conversion function, such as a _2f ^* input signal level is not affected be varied Was found to exist. One example is Is a conversion. However, α ₁ , β ₁ , α ₂ , β ₂ , K ₁ ,
K ₂ is an arbitrary constant. By using the converted signals a _f ^* and a _2f ^* as the position detection signals, the relationship described above with respect to the characteristic curve is maintained and the influence of the fluctuation of the input signal level is not exerted. However, position detection with the effect of automatic gain adjustment is possible.

The function conversion process described above is a so-called normal process. More specifically, the function conversion process is performed at the input level (a
_f, a _2f) is changed even if the output level ^{_{(a f *, a 2f *}} ) is a regular process that does not change.

That is, according to the present invention, a predetermined beam obtained from the mark portion provided on the first member is applied to the mark member or a second member having a slit, and the beam obtained from the second member is detected by a beam detector. In addition, a method of detecting the position of the first or second member by vibrating the beam and the second member relatively and detecting the detection output of the beam detector in synchronization with this vibration. In each of the signals of the fundamental component a _f and the nth harmonic component a _2f (n is an integer of 2 or more) of the output obtained by the synchronous detection, the peak point of the fundamental component a _f and the 2 harmonic component a _2f
Performing function conversion for normalization processing that eliminates input level fluctuations while maintaining the correspondence with the zero point of, and position detection is performed using the output components a _f ^* , a _2f ^* obtained by this conversion as detection signals. The position detection method is characterized by

For example, the fundamental wave component a _{f of the} output obtained by the synchronous detection
And the second harmonic component a _2f as input, and the above fundamental wave component a _f
Of performs conversion to erase the input level variation maintains a relationship between the peak point and the zeros of the second harmonic component a _2f, ^* the conversion output component obtained by a ^_f, a position detection as a detection signal to a _2f ^* It can be performed.

Example of Invention

FIG. 4 is a schematic configuration diagram showing a vibration type position detecting device used in the method of the embodiment of the present invention. The light beam emitted from the beam irradiation source including the light source 1 and the focusing lens 2 is
The object to be inspected 5 through the half mirror 3 and the objective lens 4.
(First member) is irradiated. For example, a linear alignment mark 6 is formed on the inspection object 5,
The light beam is irradiated near the mark 6. The reflected light from the mark 6 due to the light beam irradiation travels upward through the objective lens 4 and the half mirror 3, passes through the minute gap 7a of the slit plate 7 (second member), and is received by the photoelectric converter 8. It The detection signal from the photoelectric converter 8 is passed through a preamplifier 9 to a signal conversion unit 20 described later.
Is supplied to.

On the other hand, the slit plate 7 is attached to the vibrator 12, and is vibrated in a direction orthogonal to the traveling direction of the reflected light. That is, according to the oscillation output of the oscillator 11, the oscillator 12
Vibrates, and the light passing through the minute gap 7a of the slit plate 7 vibrates. The oscillation output of the oscillator 11 is also supplied to the signal converter 20 together with the vibrator 12. As will be described later, the signal conversion unit 20 detects the input detection signal in synchronization with the oscillation output, performs desired signal conversion, and supplies the conversion output to the indicator meter 13.

Now, as shown in FIG. 5, the signal conversion section 20 includes the tuning amplification circuits 21a and 21b, the synchronous detection circuits 22a and 22b, and the amplifier 2.
3a, 23b, adder 24, limiter 25 and divider 2
It is composed of 6a and 26b. That is, the detection signal input through the preamplifier 9 is supplied to two tuning amplification circuits 21a and 21b having different tuning frequencies f and 2f, respectively, and the amplified output of each circuit 21a and 21b is supplied to the synchronous detection circuits 22a and 22b. Each is supplied. The synchronous detection circuits 22a and 22b synchronously detect the amplified signals in synchronization with the oscillation output of the oscillator 11. Thus the fundamental wave component a _f detected in the synchronous detection circuit 22a outputs is outputted, the synchronous detection circuit 22b is the second harmonic component a _2f is outputted. The two detection outputs a _f and a _2f are amplifiers 23a and 2a.
Amplified by 3b and converted into α | a _f | and β | a _2f |, which are input to the adder 24. Output α from adder 24
| A _f | + β | a _2f | passes through the limiter 25 and the divider 26a,
26b. Then, from the dividers 26a and 26b, α · a _f / (α | a _f | + β | a _2f |), β · a _2f / (α | a _f
| + Β | a _2f |) is supplied to the indicating meter 13 as a new position detection signal.

Next, a position detection method using the above device will be described. First, the principle of the vibration type position detection method is the same as the conventional one, and the slit plate 7 is vibrated, and the reflected light from the mark 6 received by the photoelectric converter 8 is vibrated. Then, by synchronously detecting the detection output, the fundamental wave component a _f and the second
The harmonic component a _2f is obtained. Up to this point, the process is the same as the conventional one, and in this embodiment, this component is converted by the signal conversion unit 20 as follows. That is, in the present embodiment, attention is paid to the fact that the outputs a _f and a _2f of the f and 2f components after synchronous detection have a special relationship with respect to the position, and calculation is performed using these two outputs themselves. Conversion of. ^* This converts the output a _f after performing, is a _2f ^* will be used as a new position detection signal.

Here, the relationship between the positions of a _f , a _2f and a _f ^* , a _2f ^* will be described. FIG. 6 is a typical output characteristic curve of a _f and a _{2 f} . a _f and a _2f relationship a _f a positional shift amount is zero
Becomes zero and a _2f shows the peak value P ₂ and a _f shows the peak value P ₁
A _2f is zero at the point. Even if there is a change in the input signal level due to a change in mark contrast, the above relationship always holds, and the ratio of a _f and a _{2 f at the} same mark position is always constant. By the way, in normal position adjustment, position correction is performed so that the position detection output a _f becomes zero, but even if the position detection curve is not a _f itself, the output shows zero at the point where the position deviation amount is zero. If the position coordinate is a monotone curve in the area B between x ₂ and x ₃ , it can be used as a position detection curve. Therefore, considering the relationship between a _f and a _2f described above, α and β are set to a constant relationship, for example, α: β = | P
₂ |: | P ₁ | At this time | P ₂
| / | P ₁ | is a known quantity that is constant if the mark width and scanning amplitude are the same. Then, considering the input / output ratio, α, β
Determine. Α was thus determined, is obtained by performing the conversion of the (1) equation using the beta a _f ^*, the output characteristics of a _2f ^* is as shown in FIG. 7, speaking about a _2f ^* The output becomes zero at the point where the positional deviation is zero and the position coordinate becomes a monotonous curve in the region B between x ₂ and x ₃ . That is, it becomes possible to perform the alignment using a _f ^* as the position detection curve. By using a _f ^* and a _{2 f} ^* together, the area B
It is also possible to detect the positions of the areas A and C outside the area.

^Next, a _f ^*, a _2f ^* is (acts AGC) always unchanged for variations in the input signal level can be explain. Output a _f , a _2f
Can be considered to change in proportion to the change in the input signal level. Considering the case where a _f → K ・ a _f a _2f → K ・ a _2f (K> 0), the formula (1) is considered. Than Next, regardless of the value of K a _f ^*, a _2f ^* it can be seen that unchanged. Therefore, by using the output characteristic curve of a _f ^* for position detection, a stable position detection signal can be obtained with the effect of AGC with respect to variations in the input signal level.

Thus, according to the method of the present embodiment, it is possible to perform the vibration-type position detection having the effect of AGC with respect to the change of the input signal level without impairing the detection accuracy. In addition, since the effect of AGC can be provided rather than performing the arithmetic processing of the output after the synchronous detection, it is possible to perform highly accurate AGC that is less affected by the noise component of the input signal. Therefore, especially when used in semiconductor manufacturing equipment, the effect of AGC is also obtained against changes in the wafer's mark contrast due to differences in processing steps and changes in the input signal level due to changes in the source of the mark signal over time. It is possible to detect the position with the position, and to perform highly accurate position detection and position adjustment. Moreover, since it is not necessary to perform extra detection for knowing the change in the input signal level, the time required for alignment can be shortened,
As a result, there is an advantage that the throughput of the device can be improved.

FIG. 8 is a block diagram for explaining another embodiment. The same parts as those in FIG. 5 are designated by the same reference numerals, and detailed description thereof will be omitted. The difference between this embodiment and the embodiment described above is that instead of AGC processing of analog signals,
That is, the AGC is performed by the arithmetic processing by the digital computer. That is, the synchronous detection circuits 22a, 22
The detection outputs a _f and a _2f of b are supplied to the digital computer 30. On the computer 30 A _f ^* , a _2f ^* are obtained by performing the following calculation. Therefore, by using the output characteristic curve of a _f ^* for position detection, position detection with the same AGC effect as in the previous embodiment is possible. When digital arithmetic processing is performed here, it has a feature that the load on the hardware of the processing circuit is lighter than in the case of analog processing, but the essential idea is the same for AGC.

The present invention is not limited to the above-mentioned embodiments. In the above-described embodiment, the case where the mark pattern shape is one line which is usually used has been described. However, this mark pattern shape is formed from the line and space shown in FIG. 9 used for the alignment mark of the electron beam transfer. It may be The electron beam transfer device
The light is irradiated onto the sample to be exposed as an electron beam pattern by a photoelectric conversion mask. For alignment,
The photoelectric conversion mask has a slit-shaped photoelectric exchange mark portion, and a slit-shaped beam is applied to a sample having a metal mark. Then, the beam is deflected and vibrated to detect the X-ray from the metal mark by the X-ray detector below the mark, and the signal is processed to detect the position. In this case a _f, a is obtained by performing the conversion of the output characteristic curve of a _{2f f} ^*, the output characteristic curve of a _2f ^* is as Figure 11. As can be seen from this figure, in this example as well, the output characteristics of a _f ^* and a _2f ^* having the same tendency as in the case of the embodiment are obtained, and position detection with the same effect as in the case of the embodiment becomes possible. I understand. That is, the position detection method of the present invention can be applied as long as the mark pattern shape is one that is used for normal position detection.

Further, a slit plate for vibrating the light beam may be provided on the light incident side, that is, it may be arranged between the beam irradiation source and the mark. Further, it is also possible to use transmitted light that has passed through the mark and the sample instead of the reflected light as the detection light. The beam irradiation source may emit an electron beam instead of light. In this case, the electron beam may be deflected as the beam oscillating means. Further, as the detection beam, reflected electrons, secondary electrons or X
A line or the like may be used.

Further, the constant in the conversion can be arbitrarily changed and used, and a value that is easy to use may be determined and used in actual use. In addition, various modifications can be made without departing from the scope of the present invention.

[Brief description of drawings]

1 to 3 are respectively for explaining a conventional vibration type position detecting method. FIG. 1 is a schematic configuration diagram showing a vibration type photoelectric microscope, and FIG. 2 is a change and output of a modulated signal waveform. FIG. 3 is a signal waveform diagram showing the characteristic curve, and FIG. 3 is a signal waveform diagram showing the change of the characteristic curve when the input level changes.
4 to 7 are each for explaining the method of one embodiment of the present invention, and FIG. 4 was used for the method of the same embodiment.
FIG. 5 is a schematic block diagram showing a position detecting device adopting the AGC method, FIG. 5 is a block diagram showing a main part configuration of the device, and FIG. 6 is a signal waveform diagram showing output characteristic curves of a fundamental frequency component and a double frequency component. , FIG. 7 is a signal waveform diagram showing the output characteristic curves of the fundamental frequency component and the double frequency component subjected to AGC, FIG. 8 is a block diagram for explaining another embodiment, and FIG. 9 is line & space. FIG. 10 is a plan view showing an alignment mark pattern shape consisting of a pattern, FIG. 10 is a signal waveform diagram showing output characteristic curves of a fundamental frequency component and a double frequency component when the mark shown in FIG. 9 is used,
FIG. 11 is a signal waveform diagram showing the output characteristic curves of the fundamental frequency component and the double frequency component which have been subjected to AGC when the mark shown in FIG. 9 is used. 1 ... Light source, 2 ... Focusing lens, 3 ... Half mirror, 4 ... Objective lens, 5 ... Inspected object, 6 ... Position detection mark, 7 ...
Slit plate, 7a ... Minute gap, 8 ... Photoelectric converter, 11 ...
Oscillator, 12 ... Oscillator, 20 ... Signal converter, 21a, 2
1b ... Tuning amplification circuit, 22a, 22b ... Synchronous detection circuit,
23a, 23b ... Amplifier, 24 ... Adder, 25 ... Limiter, 26a, 26b ... Divider.

Claims (8)

[Claims] 1. A predetermined beam obtained from a mark portion provided on a first member is applied to a mark member or a second member having a slit, and the beam obtained from the second member is detected by a beam detector. In addition, a method of detecting the position of the first or second member by vibrating the beam and the second member relatively and detecting the detection output of the beam detector in synchronization with this vibration. At the peak point of the fundamental wave component a _f for each signal of the fundamental wave component a _f and the nth harmonic component a _2f (n is an integer of 2 or more) of the output obtained by the synchronous detection. Second harmonic component a
Function conversion is performed for normalization processing that eliminates input level fluctuations while maintaining the correspondence with the zero of _2f , and position detection is performed using the output components a _f ^* and a _2f ^* obtained by this conversion as detection signals. A position detection method characterized by the above. 2. The position detecting method according to claim 1, wherein the conversion means for eliminating the fluctuation of the input level is the conversion shown by the following equation. However, α ₁ , α ₂ , β ₁ , β ₂ , K ₁ , and K ₂ are arbitrary constants. 3. As conversion means for eliminating the fluctuation of the input level, each peak value P ₁ , P ₂ of the output characteristic curves of a _f and a _2f
P ₂ is The position detecting method according to claim 2, characterized in that the following relationship is held. 4. The position detecting method according to claim 1, wherein the mark provided on the first member is irradiated with a beam from a beam irradiation source. 5. The first member is a photoelectric exchange mask, and the mark member of the second member is X-ray irradiated with an electron beam.
The beam detector is made of a material that produces a line, and the beam detector is
The position detecting method according to claim 4, which is a line detector. 6. The beam irradiation source emits light, and the beam detector detects light passing through the mark portion or light reflected by the mark portion. The position detecting method according to the first item of the range. 7. The means for vibrating the beam and the second member relative to each other vibrates a slit plate as the second member in a direction orthogonal to the traveling direction of the light. The position detection method according to claim 5. 8. The beam irradiation source is for irradiating an electron beam, and the beam detector is for detecting backscattered electrons or secondary electrons or X-rays from the mark portion. The position detecting method according to claim 1.